Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS3621397 A
Publication typeGrant
Publication dateNov 16, 1971
Filing dateJul 22, 1969
Priority dateJul 22, 1968
Also published asDE1937132A1, DE1937132B2
Publication numberUS 3621397 A, US 3621397A, US-A-3621397, US3621397 A, US3621397A
InventorsMasayoshi Murotani, Ryoichi Tanaka, Keiji Tachikawa, Tadao Shimamura
Original AssigneeNippon Telegraph & Telephone, Nippon Electric Co
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Pcm transmission system
US 3621397 A
Abstract  available in
Images(8)
Previous page
Next page
Claims  available in
Description  (OCR text may contain errors)

United States Patent Inventors Masayoehl Murotenl Kanagawa; Ryoichi Tomb, Tokyo; Kelli Tachlkawn, Chiba; Tadao Shlmamun, Tokyo, all oil Japan Appl. No. 843,583 Filed Jilly 22, 1969 Patented Nov. 16, 1971 Assignees Nippon Telegraph & Telephone Public Corporation y 3 Nippon Electric Company, Limited Tokyo, Japan Priority July 22, 1968 Japan 43/52062 PCM TRANSMISSION SYSTEM 14 Claims, 27 Drawing Figs. US. Cl 325/38, 179/1555, 178/66 lat. Cl. H031: 13/22, l-l04b 1/66 FleldotSear-ch l79/15BV,

l5 BW, 15 AP, 15.55; l78/DIG. 3, 66, 67, 68;

Primary Examiner-Benedict V. Safoulrek Anorney-Marn and Jangarathis ABSTRACT: A transmission system for a television signal represented by a pulse code modulation signal compressed in bandwidth, comprising a transmitter for dividing the television signal into a preselected number of samples indicated by a corresponding number of first sets of parallel pulse code modulation pulses which are converted into a preselected number of second sets of parallel pulse code modulation pulses containing 0" and "1 logic signals to indicate slow and rapid changes in the television signal magnitude at successive even-number signal samples; the preselected second number being smaller than the preselected first number; the second parallel pulse sets being transmitted as series pulse code modulation pulses; and a receiver for converting the received series pulses into a third set of parallel pulse code modulation pulses including the 0" and l "logic signals and corresponding to the transmitter second pulse sets, and translating the receiver third pulse sets into a reproduction of the transmitter 325/38, 38 13 television signal.

3! 4| SAMPLER BAND jg PARALLEL 34 :4

ENCODER 5 L COM 5 SER'ES T" 5| 52 36 PRESSOR cowv. 49

55 SIGNAL CLOCK 56 r 5? SOURCE GEN. 53

J' 4 67 EW- [EDMU- I GHAR ATION IFER-I PAR- IDELAYI' I ENCE AT R DISCRIMINATOR L J L EU J l THRESH- 6 OLD (h) 2 This invention relates to a system for transmitting a P'CM signal with band compression and, more particularly, to a system of the kind applicable to the transmission of television picture signals.

in general, a television picture signal has meat redundancy. n the other hand, it is known that PCM transmission is highly insusceptible to noise but requires a considerably wider frequency hand than other types of transmission.

An approach to compressing of the bandwidth of a teievision signal or other approximately repetitive signals is to utilize, according to information theory, the fact that the conditional entropy of the signal is smaller than the primary entropy. More particularly, the amount of the information to be actually transmitted is reduced by using the correlation which is positive for the signal portions spaced by the period of the approximate repetition, such as for the signal portions representing the respective picture elements of the consecu tive scanning lines and also of the consecutive frames (Peter Elias, Predictive Coding, IRE Transaction on information Theory, Mar. i955, pp. 16-33; Robert E. Draham, Predio tive Quantizing of Television Signals," IRE Wesson Convention Record, Aug. 1958, pp. l47-l56). The system according to this theory however, requires a highly complicated memory device.

Another approach is to resort to visual psychophysics. This does not fall in the category of band compression in the strict sense as regards the reproducibility of the information contained in the original signal but provides band compression when the information receptor (the human visual nerves in the case of the television signal) is deemed as a part of the information transmission system. The band compression of this sort is based on the fact that the discriminating ability of the human eyes depends on the rate or speed of variation of the luminance levels. More particularly, slow change in the luminance level of successive picture enables the human eyes to discriminate minute difference in the luminance level, whereas the discriminating ability decreases when the luminance level varies quickly. The band compression is also based on the fact that the human visual sense responds to the contour of an object or pattern as a whole and to the luminosity of the internal areas.

An example of this kind of approach is the Synthetic Highs system proposed by W. IF. Schreiber ("Synthetic Highs-An Experimental TV Bandwidth Reduction System, Journal ofthe SMPTE, Volume 68, Aug. 1959, pp. 525-537), which comprises a filter for dividing the picture signal into a lower and a higher frequency component, means for transmitting the lower component as an analogue signal without any modification, and means for transmitting the higher component after having encoded the same with comparatively rough quantization. The higher component results from the border lines of the patterns contained in a picture. The number of such border lines is very small in general. in other words, the amount of the information concerning these border lines is small as compared with the amount of the information regarding the whole picture. It is therefore possible, by coarsely quantizing such information and transmitting the encoded information together with the information concerning the positions of such border lines, to transmit the higher frequency component at a reduced speed as compared with the speed required to transmit the whole picture signal as it occurs.

in another example, the lower frequency component is subjected to fine quantization while the higher frequency com ponent is subjected to coarse quantization (IE. it. liretzner, Reduced-Alphabet Representative of Television Signals," IRE Convention Records, Part ill, 1956, pp. l4(Ll47). in this system, the sampling frequency is high and the number of bits in a codeword is small for the higher component, while the sampling frequency and the number of bits in a code word are low and large, respectively, for the lower component. Transhigher components and two channels for transmitting these components.

SUMY OF THE INVENTION It is therefore a general object of this invention to provide a simplified system for transmitting a lPCM signal with band comprmsion.

It is a specific object to provide a PCM signal with band compression specifically applicable to the transmission of a television signal.

The system of this invention as applied to the transmission of a television picture signal is based on visual psychophysics. Study of psychophysics has revealed that there is an upper limit in the amount of the information acceptable by the human visual sense during a unit time, with the result that the amount of the luminance information decreases with the increase in the amount of the spatial or geometrical information and vice versa. It follows therefore that the transmission of the entire information extending beyond the upper limit is waste ful and that the information should be transmitted at a speed below the speed corresponding to the upper limit.

According to a generalized aspect of this invention, there is provided:

in a lPCM transmission system for data variable with time,

and represented by PCM codewords of a first kind, respectively, said first-kind codewords appearing at a rate of p per unit time where p is a given positive number,

a transmitter comprising:

means responsive to q consecutive ones of said data where q is an integer greater than two, for sensing the speed of variation of such consecutive data to produce a variationrepresenting signal element, having the form of at least one digit of PCM codewords of a second kind, said second-kind codewords appearing at a rate of p/r per said unit time where r is an integer greater than one, said signal element representing at least two discrete values which correspond to the respective degrees of said speed, and

means responsive to said first-kind codewords and said signal elements for producing the second-kind codewords, each said second-kind codeword standing for r consecutive first-kind codewords, said second-kind codewords containing at first prescribed bit positions said signal elements, respectively, each said second-kind codeword further containing at a second bit position the codes of the more significant digits of a preselected one of said r first-kind codewords, each said second-kind codeword still further containing at the remaining bit position the codes of the remaining digits of said preselected firstkind codeword whenever the signal element contained therein shows slow variation of the data and the codes of the more significant digits of the predetermined at least one of said r consecutive first-kind codewords except said preselected one whenever the signal element contained therein shows quicker variation.

The data as called herein may be represented by either a digital signal or by samples derived from an analogue signal. In the latter case, p is the samples per unit time. The variationrepresenting signal-producing means may be supplied at a time time with the analogue signal portion covering either r consecutive samples or r first-kind PCM codewords. It is now assumed that r is 4, that the speed is divided into 3, that two codewords are predetermined from four consecutive codewords for medium speed, and that four codewords are predetermined similarly for high speed. The preselected firstkind codeword may be the first, the second, the third, or the fourth of the four codewords. Depending on which of the four codewords is preselected, the predetennined two codewords may be either of a set of the first and the third codewords or of a set of the second and the fourth codewords. It is understood that seven combinations of codewords are possible in which one, two and four correspond to the slow, the medium and the high speeds, respectively. Under the circumstances, it is necessary to provide seven discrete values for the signal elements.

For example, an analogue signal is sampled at a sampling frequency of 10 MHz and encoded into eight-bit PCM codewords. The variation-representing signal element is produced with reference to four consecutive samples to represent whether the variation is rapid, moderate, or slow. When the variation is rapid, the more significant two digits of each codeword are transmitted so that two bits are transmitted per datum sampled at 10 MHz. repetition frequency. When the variation is moderate, the more significant four digits of every other codeword are transmitted so that four bits are transmitted per datum sampled at MHz. When the variation is slow, the full digits are transmitted for only one preselected codeword in each set of four consecutive codewords so that eight bits are transmitted per datum sampled at 2.5 MHz. The variation-representing signal element has three bits to represent the seven combinations when the binary code is used for the PCM signal. Under the circumstances, the original PCM signal which would have been transmitted at a speed of 80 megabits per second is transmitted as a modified PCM signal which equivalently contains 1 l-bit codewords for the data sampled at 2.5 MHz. and is transmitted at a speed of 27.5 megabits per second. TI-Ie ratio of compression is about one-third.

In another example, the variation-representing signal element has one bit representing the slow and the rapid variations. When the variation is rapid, four bits are transmitted at a repetition frequency of MHz. When the variation is slow, full eight digits are transmitted at a second repetition frequency of 5 MHz. The transmission speed for the modified PCM signal is thus (8+l 5=45 megabits per second. This provides compression of about one-half.

According to this invention, transmission of the whole PCM signal is carried out at a substantially constant speed with only one encoder and without a memory device which has been indispensable in the output circuit to attain the nearly equal speed. This invention thus provides a much simplified system of the kind.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a transmitter of the system of this invention;

FIGS. 2A and 2B show signals for explaining the operation of a variation discriminator in the transmitter;

FIG. 3 shows a portion of the original PCM signal;

FIG. 4 shows a portion of the band-compressed PCM signal;

FIG. 5 is a block diagram of a difference deriver;

FIG. 6 shows a portion of the analogue signal for explaining the operation of the variation discriminator;

FIG. 7 is a block diagram of another difference deriver;

FIG. 8 is a circuit diagram of the variation discriminator;

FIG. 9 is a circuit diagram of a band compressor in the transmitter;

FIG. 10 shows portions of various signals for explaining the operation of the band compressor;

FIG. 1 l is a block diagram of a receiver of the system of this invention;

FIG. 12 shows a portion of the received band-compressed PCM signal;

FIGS. 13 through 15 show portions of various signals in the receiver;

FIG. 16 is a circuit diagram of a time rearranging circuit in the receiver;

FIG. 17 shows portions of various signals for explaining the operation of the time-rearranging circuit;

FIG. 18 is a circuit diagram of an adder circuit in the receiver;

FIG. 19 shows portions of various signals for explaining the operation of the adder circuit;

FIG. 20 is a circuit diagram of a combining circuit in the receiver;

FIG. 21 shows portions of various signals for explaining the operation of the combining circuit;

FIG. 22 is a block diagram of another receiver of the system of the invention; and

FIGS. 23A-23D show portions of various signals for explaining the operation of the receiver in FIG. 22.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Referring to FIGS. 1 through 4 inclusive, it is presumed that each sample S, (i is an integer) shown in FIG. 2A of an analogue signal 30 is encoded in a transmitter of this invention into six-bit parallel PCM binary pulses 31-36 shown in FIG. 3, that the original PCM pulses 31-36 are band-compressed into seven-bit parallel PCM output pulses 41-47 shown in FIG. 4, which may be subsequently converted into a series PCM output signal 49, and that the PCM output pulses 41-47 or signal 49 is band-compressed, in efi'ect, from six bits per sample into 3.5 bits per sample. a

The transmitter comprises a signal source 51 of the analogue signal 30 and a clock generator 52 generating sampling pulses 54 of a sampling period t, timing pulses 55 of a repetition period 2!, and first clock pulses 56 having a repetition period of I and a common pulse width of r/2. The clock generator 52 further generates second, third, and fourth clock pulse trains 57, 58, and 59 having a common repetition period of 2! and a common pulse width of !/2, these clock pulse trains 57, 58, and 59 being shifted in time from the first clock pulse train 56 by the respective amounts mentioned hereunder.

The transmitter further comprises a sampler-encoder 61 for encoding each sample S, derived therein by the sampling pulses 54 from the analogue signal 30 supplied thereto, into a set of six parallel PCM pulses 31-36 in the known manner. Thus, the samples 8,, 8,, S,, are represented by the successive of PCM Pulms n n my re), (an n ss) ---i n, a a shown in FIG. 3, each bit a (i is an integer) being either logic l or 0" in case the pCM signal is of the binary form, the first bit 0,, in each set being assumed to represent the most significant digit of a set of the PCM pulses 31-36 for the ith sample 8,.

The transmitter still further comprises a source 62 of a threshold signal h (FIG. 2B) and a variation discriminator 63 including, as will be described later in detail, a difierence deriver 64 with linear, parabolic, or other interpolation supplied with the analogue signal 30, a comparator 65 supplied with the output x of the deriver 64, the threshold signal h, and the timing pulses 55, and a delay circuit 66 for delaying the output of the comparator 65 to provide logic "1 and 0" variation-representing pulses 67 (FIG. 3). By virtue of the timing pulses 55, the difference deriver 64 with linear interpolation, for example, develops in effect the samples 8,, produces with reference to the odd-numbered samples S and S,,,., for reference, even-numbered calculated samples 8,, (FIG. 2A) and derives, in principle, that difference signal 2 (FIG. 2B) representing the difi'erence between the true and the calculated even-numbered samples 8,, and 8,, which represents the rate or speed of variation of the analogue signal 30 or the derived PCM pulses 31-36 at the portion of the sampling point T for the even-numbered sample 8,,. The comparator 65 produces logic 1 and 0" pulses of a common pulse width of twice as long as the sampling period I when the difference signal e is greater and smaller than the threshold signal h, respectivelyfin absolute value. The delay circuit 66 gives delay to the 1" and 0" pulses so that each of the variationrepresenting pulses 67 may be in time registered with those two consecutive PCM pulse sets of the PCM pulses 31-36, the speed of variation at which portion is represented by the particular variation-representing pulse 67. For example, the variation-reprwenting pulses 67 vary from 6 to "ll," as shown in F116. 3, in time coincidence with the leading edges of the PCM pulses a a and a for the fourth sample 8,, at which portion the analogue signal 36 varies rapidly and returns to 6" simultaneously with the trailing edges of the PCM pulses a a and a for the seventh sample S preceding the following sample 8,; at which portion the speed of variation is slow.

The transmitter further comprises a band compressor 66 for deriving, in the manner mentioned below in detail, seven-bit band-compressed PCM output pulses 61-67 from the original PCM pulses 311-36 and the variation-representing pulses 67 by using the first through the fourth clock pulses 66-66. As shown in FIG. 6, the successive sets of the PCM output pulses all-47 consist of (a a a 6), a a a a a ll"), The first bit ra in each set represents the most significant digit of the PCM output pulses all-66 for the odd-numbered sample S while the seventh bit represents the speed of variation. it will be seen that the more significant three digits a a and a "3 of the original IPCM pulses 311-36 for the reference samples S are always used as the corresponding digits of the PCM output pulses 41-67 with the pulse width doubled and with the leading edges delayed by a sampling period 1 relative to the variation-representing pulses 67 or the seventh-bit pulses. When the seventh bits are 6" and "11, the less significant three digits "(en-1m mt-us and rea-ms of the original PCM Pulses 3ll36 for the reference sample S with the leading edges delayed by the sampling period t and more significant three digits a 11, 4, and dun-.213 of the original PCM pulses 311-36 for the even-numbered true sample S without the delay are used, respectively, as the less significant three digits of the PCM output pulses 4li67, with the pulse width being doubled. in this manner, the band compressor 66 compresses the original PCM pulses 311-36 containing six bits of information per sample into the PCM output pulses ll-i7 containing 7/2 bits per sample.

The transmitter may further comprise a parallel-series converter 69 for converting the band-compressed PCM output pulses M 67 into a series PCM output signal 69.

Referring further to FIG. 2a and also to H618. 5 and 6, the difference deriver M with linear interpolation comprises first and a second ideal delay lines 7i and 72, respectively, each having a delay time of the sampling period I, At the moment the third sample S of a set of three samples S S, and 8 reaches the input of the first delay line 7 l, the second and the first samples S, and 8 reach the outputs of the first and the second delay lines 711 and 72, respectively. The deriver 64 further comprises an adder 73 for deriving the sum of the samples S and S and an amplifier 7a with a gain of one-half for dividing the sum by two to provide the calculated sample 8,. The deriver 66 still further comprises a subtractor 755 for subtracting the calculated sample S" from the corresponding true sample S, to give the difference signal e I given by the following equation:

Referring further to H6. 6 and also to H6. '7, a difference deriver 6-6 with parabolic interpolation comprises a first delay line 76 having a time delay of 2! and a second and a third delay lines 77 and 76, respectively, each providing a time delay of r. At the moment the sample S reaches the input of the first delay line 76, the samples S 8,, and 5 reach the outputs of the first, the second, and the third delay lines 76, 77, and 78, respectively. The deriver M further comprises first, second, and third amplifiers 61, 62, and 63 having the gains of minus one-eighth (a combination of an inverter and an attenuator), six-eighths, and three-eighths, respectively, for the input of the first delay line 76 and the outputs of the second and the third delay lines '77 and 76, respectively. The deriver 64 still further comprises an adder 66 for deriving the sum of the amplifier outputs to provide the calculated sample S, and a subtractor 66 for subtracting the sum from the true sample S, to deliver the difference signal e, given in this case by the following relation:

Referring to FIG. 6, the difierence deriver 64 with linear interpolation comprises an analogue signal input terminal of the variation discriminator 63, buffer amplifier 61 of unit gain for the analogue signal 30, and a delay circuit 64a including first and second delay lines, 62 and 63, respectively. Each of the delay lines 62 and 63 has a delay time of the sampling period r and may be either a lumped-constant delay network or a distributed-constant delay line (for example, a coaxial cable). The delay circuit 64a is terminated by a resistor 84 whose resistance is equal to the characteristic impedance of the delay lines 62 and 63. By virtue of the timing pulses 55 supplied to a timing pulse input terminal 85 of the discriminator 63, signals appearing at the tap points of the delay circuit correspond to the samples S S and sg qgq, respectively, provided that the delay lines are ideal and have no insertion loss. These signals are applied to first, second, and third bufi'er amplifiers 66, 67, and 88, respectively. Each of the first and the third amplifiers 66 and 86 has a unit voltage gain, while the second amplifier 67 has minus Twice the unit voltage gain. The outputs of the respective amplifiers 86, 87, and 88 are led to a resistor adder consisting of three resistors 91, 92, and 93, each of which has the same resistance. The output signal x of the adder is two-thirds of the difference signal e, because of the following equality:

fi zrmn swf ufl/ 3 2I'1BI+ 2l+r)/ 2l] With an actual delay line which exhibits insertion loss, it is necessary to change either the gains of the buffer amplifiers 66, 67, and 68 or the resistances of the resistors 9i, 92, and 93 or both so as to compensate for the insertion loss of the delay lines 62 and 63.

The comparator 65 comprises a voltage amplifier 99 of a gain G selected in the manner mentioned below for amplifying the difference deriver output signal x to produce an amplified output y and a comparator unit 166 having first and second input transistors llilll and 1162, respectively, and first and second paired transistor 1163 and 1104, respectively. The aim plified output y is supplied to the bases of the first and the second input transistors 1611 and 1162. The emitters of the first input and paired transistors i and 1163 are connected with a biassing source 1165 of the biassing voltage V (a negative voltage) via a first biassing resistor 166. The emitters of the second input and paired transistors 1162 and 1104 are also connected with the biassing source K05 via a second biassing resistor 1167. Similarly, the collectors of the second input and the first paired transistors 1162 and 1103 are connected, via a load resistor 1166 and a diode 109, with first and second power supplies lllll and 112, respectively, of the first and the second power voltages V and V respectively. Furthermore, the collectors of the first input and the second paired transistors lltlll and 166 are connected with the second power supply 112. The relation between the first and the second power voltages V and V is given by the following inequalities:

and

CCZ IW CCI IIOB) where V and V are the voltage drop across the resistor 166 and the forward drop of the diode 169, respectively. The bases of the paired transistors 1103 and TM are supplied with reference voltages h and Hi, respectively, by the threshold signal source 62. Each of the transistor pairs 1101 and I163 or 162 and 1166 forms a differential amplifier.

If y k, the paired transistors 103 and 104 are conductive. lf h y +h, the first input and the second paired transistors i611 and l and the first paired and the second input transistors 1163 and 102 are conductive and nonconductive, respectively so that no current flows through the load resistor 1166. in this case, the discrimination output z of the compara tor unit 100 is equal to the first power voltage V lf +h y, the input transistors and 102 are conductive. It follows therefore that in case the amplified output y is either lower than h or higher than +h, current flows through the load resistor 108. Under the circumstances, the discrimination output 2 is clamped to the voltage V -V Thus, the comparator unit 100 discriminates whether the amplified output y is greater or smaller than the reference voltage or the threshold signal I: in absolute value.

In practice, the comparator unit 100 may not correctly discriminate the amplified output y when it is nearly equal to the threshold signal It in absolute value. This is due to the imbalance in each differential amplifier. Let it be assumed that voltages a and +a supplied to the respective bases of the first and the second input transistors 101 and 102 balance the differential amplifiers. Furthermore, let it be recalled that the threshold signal I: is selected for the difference signal e. lnasmuch as the difference deriver output signal 1: is equal to 2e/3, either the gain G of the voltage amplifier 99 should be equal to 3a/( 2):) or the reference voltage should be two-thirds of the threshold signal In with the gain G being set at the unit.

The comparator 65 further comprises a logic circuit assembly in turn comprising an AND/NAND gate 115 to which the discrimination output 2 of the comparator unit 100 is supplied, after the level thereof is adjusted by a slicer 116 to a level suitable for the AND/NAND gate 1 15. The AND/NAND gate 1 produces the input signal as it stands at an AND output terminal 117 and an inverted signal at a NAND output terminal 118. When the discrimination output 2 is Vcci, the outputs at the AND and the NAND output terminals 117 and 118 are logic "1 and 0, respectively. When it is equal to V V such outputs are 0" and 1," respectively. These binary codes are delivered to a flip-flop 119, which supplies the NAND output to the delay circuit 66 in accordance with the timing pulses 55. It should be mentioned here that the sampler-encoder 61 produces the bit-parallel PCM pulses 31-36 with a certain time delay relative to the supplied analogue signal 30. The delay circuit 66 is used to make the leading edges of the variation-representing pulses 67 coincide with the leading edges of the PCM pulses 31-36 of every other sample.

Referring to FIGS. 9 and 10, the band compressor 68 comprises first, fourth, and seventh channels 121, 124, and 127, respectively. The first channel 121 is coupled with the fourth channel 124. Likewise, the second and the third channels (not shown), similar in construction to the first channel 121, are coupled with the fifth and the sixth channels (not shown), similar to the fourth channel 124, respectively. The seventh channel 127 is coupled directly with the fourth, the fifth, and the sixth channels 124, All channels 121, 124, and 127 comprise input AND/NAND gates 131, 134, and 137 supplied with the PCM original pulses 31-36 and the variation-representing pulses 67, respectively. When supplied with 0" and 1" pulses, the AND/NAND gate produces 0 and 1" pulses at the NAND output terminal, respectively. This first through the sixth channels 121, 124, comprise a first set of six flip-flops 141, 144, supplied with the AND and the NAND outputs of the first-channel through the sixth-channel AND/NAND gates 131, 134, (AND outputs of the first-channel and the fourth-channel AND/NAND gates 131 and 134 are shown in FIG. 10), respectively, and with the first clock pulses 56 (shown in FIG. 10) via an inverter 149. TI-Iese six channels 121, 124, further comprise a second set of flip-flops 151, 154, which receive both AND and NAND outputs of the first-set flip-flops 141, 144, respectively and the first clock pulses 56 as they stand. The AND outputs (those of the second-set first-channel and fourth-channel flip-flops 151 and 154 are shown in FIG. 10) are six-bit parallel PCM pulses delayed by a sampling period 1 relative to the corresponding outputs of the input AND/NAND gates 131, 134, The fourth channel 124 comprises first and second NAND gates 158 and 159 respectively, supplied with the AND outputs of the first-channel input AND/NAND gate 131 and of the fourth-channel second-set flip-flop 154 together with the AND and the NAND outputs of the seventh-channel input AND/NAND gate 137, respectively. The fourth channel 124 further comprises an intennediate AND/NAND gate supplied with the outputs of the NAND gates 158 and 159. The fifth and the sixth channels comprise similar gates (not shown). When the variation-representing pulses 67 are l," the gates 158-160 produce at the NAND output of the intermediate AND/NAND gate 160 the AND output of the first-channel input AND/NAND gate 131 with a time delay inherent to the gates 158-160. While the same are 0, the gates 158-160 produce at the NAND output terminal the AND output of the fourth-channel input AND/NAND gate 134 with a time delay equal to the sum of the sampling period I and the delay inherent to the gates 158-160. In general, the operation of the gates 158-160 is represented by a logic relation:

where a is the output of the intennediate AND/NAND gate 160 and d, represents the AND outputs of the first-channel through the third-channel input AND/NAND gates 131, the fourth-channel through the sixth-channel second-set flip-flops 154, and the seventh-channel input AND/NAND gate 137. All channels 121, 124, and 127 further comprise a third set of flip-flops 161, 164, and 167 supplied with the outputs of the first-channel through the third-channel second-set flip-flops 151, the fourth-channel through the sixth-channel intermediate AND/NAND gates 160, and the seventh-channel input AND/NAND gate 137, respectively, and with the second clock pulses 57. The AND outputs of these flip-flops 161, of the first through the third channels 121, the NAND outputs of these flip-flops 164, of the fourth through the sixth channels 124, and the AND output of the flip-flop 167 of the seventh channel 127 are, as exemplified by the outputs of the first-channel and the fourth-channel flip-flops 161 and 164 in FIG. 10, band-compressed pulses whose leading edges are substantially coincident with the leading edges of the second clock pulses 57. All channels 121, 124, and 127 still further comprise a fourth and a fifth set offlip-flops 171, 174, and 177 and 181, 184, and 187 supplied, as depicted, with both AND and NAND outputs of the corresponding flip-flops of the preceding stages and with the third and the fourth clock pulses 58 and 59. The AND outputs of the fifth-set flip-flops 181, of the first through the third channels 121, the NAND outputs of such flip-flops 184, of the fourth through the sixth channels 124, and the AND output of the flip-flop 187 of the seventh channel 127 are the desired band-compressed PCM output pulses 41-47, some of which are reproduced in FIG. 10. The flipfiops of the fourth and the fifth sets are provided with a view to putting the PCM output pulses 41-47 in best order.

It should be mentioned here that a low-pass filter (not shown) may be interposed between the signal source 51 on the one hand and the sampler-encoder 61 and the variation discriminator 63 on the other hand. Also, an already-sampled analogue signal may be supplied to an encoder and a variation discriminator (corresponding to the sampler-encoder 61 and the variation discriminator 63, respectively). Alternatively, a variation discriminator, similar to the variation discriminator 63, may produce the variation-representing pulses 67 from the original PCM pulses 31-36.

Referring to FIGS 11 through 15 inclusive, a receiver of this invention to be coupled with the transmitter illustrated in FIG. 1 comprises an input terminal 200 for the band-compressed series PCM pulses, a clock generator 201 for producing a synchronism with the bits and frames of the received series PCM pulses, timing pulses 202 having the repetition period of twice the sampling period 2!, first clock pulses 203 having a repetition period 2! and a common pulse width t, and second and third clock pulses 204 and 205 having a common repetition period t. The clock generator 201 further produces fourth clock pulses 266, identical in wavefonn to the first cloclt pulses 263 but shifted therefrom by an amount which will become clear later. The clock generator 261 still further produces fifth and sixth clock pulses 267 and 266 having a repetition period t.

The receiver further comprises a series-parallel converter 210 for converting with reference to the timing pulses 262 the series PCM pulses into reproductions 211-217 (FIG. 12) of the band-compressed parallel PCM output pulses 61- 17 of the transmitter, and a time rearrange circuit 226 for rearranging the band-compressed PCM pulses 211-217 into time rearranged lPCM pulses 221-226 (FIG. 13) with reference to the first through the third clock pulses 263-265. In case the seventh-bit pulses 217 of the band-compressed PCM pulses 211-217 are logic 6, the rearrange circuit 226 produces as the time-rearranged PCM pulses 221-226 the PCM pulses c a and a for a first period corresponding to the earlier half of each repetition period of the first clock pulses 203 and logic pulses for a second period corresponding to the latter half. In case the seventh bits are 1," the rearrange circuit 226 produces as the more significant three digits of the time-rearranged PCM pulses 221-223 the PCM pulses d a and gflqn for the first period and the PCM pulses a a and a for the second period. In this case, the rearrange circuit 226 further produces logic 1" pulses as the fourth-bit time rearranged PClVll pulses 221 and logic 6" pulses as the fifth-bit and the sixth-bit pulses 225 and 226.

The receiver still further comprises first register 236 for giving a delay of one bit time to the time rearranged lPCM pulses 221-226 to produce first delayed time rearranged lPCM pulses 231-236 and a second register 246 similarly producing second delayed time rearranged PCM pulses 241-246 further delayed by one bit time. The registers 236 and 2 16 may be composed of flip-flops.

The receiver further comprises a mean value deriver 256 for digitally adding the time-rearranged and the second delayed PCM pulses 221-226 and 241-266 and for dividing the digital sum by two (by shifting, when the PCM pulses are of the binary code, the digit of the sum by one bit towards the most significant digit) to produce mean PCM pulses 251-256 (FIG. 14). It should be noted here that linear interpolation is carried out in the transmitter, that the variation-representing seventhbit pulses 217 of the band-compressed PCM pulses 211-217 and the fourth clock pulses 266 are supplied to the mean value driver 256 to adjust the timing of the produced mean PCM pulses 251-256 relative to the first delayed PCM pulses 231-236 and to suppress the mean PCM pulses a a a 1, 6," and 6" and the like which are present in the timerearranged PCM pulses 221-226 when the variation of the original data is rapid, and that 1, 6," and 6" are selected for the less-significant three digits of the time-rearranged PCM pulses 221-226 when the seventh-bit pulses 217 are 1 because a a a 1," 6, and "6 represent the mean value of a a a 1", 1", and 1" and 0, 0 a 6," 6, and "6 and consequently minimize on the average the error introduced by omission of the three less significant digits of the original PCM pulses.

The receiver further comprises a signal combiner 266 for superposing the mean PCM pulses 251-256 on the first delayed PCM pulses 231-236 in pertinent time relation pro vided by the fifth and the sixth clock pulses 267 and 266 to produce approximately reproduced PCM pulses 261-266 (FIG. 15), and a utilization circuit 276 for utilizing the reproduced PCM pulses 261-266.

Referring now to F163. 16 and 17 the rearrange circuit 226 comprises a first through a seventh channels 361, 362, and 307 of which the third and the sixth channels are not depicted. The fourth through the sixth channels 366, 365, are coupled with the first through the third channels 361, 362, respectively, while the seventh channel 367 is coupled directly with the fourth through the sixth channels 364, 365, The first through the third channels 361, 362, are similar in construction. Likewise, the sixth channel is substantially same as the fifth channel 365. All channels 361, 362, and 367 comprise input NAND gates 311, 312, and 317 supplied with the band-compressed PCM pulses 2111-217, respectively (the first, the fourth, and the seventh bits are reproduced in FIG. 17). Each of the first-channel through the third-channel NAND gates 311, 312, also receives the first clock pulses 263 (1 16. 17), while each of the fourth-channel through the sixth-channel NAND gates 314, 315, receives the first clock pulses 263 through an inverter 319 and the seventh-channel PCM pulses 217. The seventh-channel NAND gate 317 receives the seventh-channel PCM pulses 217. The fourth channel 3 comprises an intermediate NAND gate 320 supplied with the fourth-channel PCM pulses 214 and the first clock pulses 263. The first through the sixth channels 301, 362, comprise AND/NAND gates 32], 322, respectively. The first-channel AND/NAND gate 321 is supplied with the outputs of the input NAND gates 31 1 and 314 of the channel 361 and the coupled fourth channel 304. The fourth-channel AND/NANlD gate 324 receives the output of the intermediate NAND gate 326 and the output of the seventh-channel NAND gate 317. The fifth-channel AND/NAND gate 325 receives the band-compressed PCM pulses 215 supplied to the channel, 365 the output of the seventh-channel NAND gate 317, and the first clock pulses 263. The NAND output a (F16. 17) of the first-channel AND/NAND gate 321 is given by the following logic relation:

where C and d, represent the first clock pulses 263 and the band-compressed PCM pulses 211-217, respectively. This relation shows that 2 is a t at the earlier half of each repetition period of the first clock pulses 263 and that, at the later half, alMND is a and 0" when a is 1" and 6, respectively. Likewise, the NAND output 0 (FIG. 17) of the fourth-channel AND/NAND gate 324 is given by the following logic relation:

from which it is apparent that a is ar and 6" at the earlier half and the later half of each repetition period of the first clock pulses 263, respectively, when a is 6 and that a is 1" while a is Similarly, the AND output a (FIG. 17) of the fifth-channel AND/NAND gate 325 is given by the following logic relation:

QEAND=QEIAOIAEF which means that a assumes the values of (n and 6" at the earlier half and the later half of the repetition period of the first clock pulses 263, respectively, when a is 6 and that 11 assumes 6" when a(2i.- .1)7' is 1". The first through the sixth channels 361, 362, further comprise flip-flops 331, 332, of a first set supplied with the AND and the NAND outputs of the AND/NAND gates 321, 322, of the corresponding channels 301, 362, and with the second clock pulses 266 (FIG. 17), and flip-flops 341, 342, of a second set supplied with the outputs of the preceding flipflops 331, 332, respectively, and with the third clock pulses 265 (FIG. 1'7 The rearrange circuit 226 thus produces the time-rearranged PCM pulses 221-226 which are put in order by the repeated readout operation carried out at the flip-flops 331, 332, 361, 342, by the second and the third clock pulses 266 and 265.

Referring to FIGS. 16 and 19, the mean value deriver 250 comprises six full adders 351-356 for carrying out the digital addition of the time rearranged and the second delayed PCM pulses 221-226 and 241-246. More particularly, the sixth adder 356 for the least significant digit receives the sixth-bit time-rearranged and second delayed PCM pulses 226 and 246 to produce 1" or carry pulses. The fifth adder 355 receives the fifth-digit pulses 225 and 245 and the carry pulses of the sixth adder 356 to produce similar carry pulses and 1 or "0 sum pulses. Likewise, each of the remaining adders 351-354 produces the carry pulses and the sum pulses of the corresponding digit. The corresponding pulses of the carry pulses delivered from the first adder 351 and the sum pulses delivered from the first through the fifth adders 351-355 are the result of the addition of a set time-rearranged PCM pulses 11 a and a corresponding set of the second delayed PCM pulses a a the decimal point of the result being shifted by one bit towards the most significant digit. Thus, these outputs of the adders 351-355, if properly timed, represent the mean value of the consecutive two sets of the time rearranged PCM pulses 221-226 (FIG. 19).

The mean value deriver 250 further comprises a timing circuit 360. The timing circuit 360 in turn comprises a NAND gate 361 supplied with the fourth clock pulses 206 (FIG. 19), an AND/NAND gate 362 supplied with the seventh-bit pulses 217 (reproduced in FIG. 19) of the band-compressed PCM pulses 211-217, a first flip-flop 366 supplied with the AND and the NAND outputs of the AND/NAND gate 362 and stepped by the output (FIG. 19) of the NAND gate 361, a second flip-flop 367 supplied with the AND and the NAND outputs of the first flip-flop 366 and stepped by the fourth clock pulses 206, and a timing output AND gate 369 supplied with the output of the NAND gate 361 and the NAND output H6. 19) of the second flip-flop 367.

The mean value deriver 250 still further comprises a first through a sixth output AND gate 371, 372, all supplied with the output (FIG. 19) of the timing output AND gate 369. The output AND gates 371, 372, are further supplied with the corresponding digits of the mean value of the consecutive two sets of the time rearranged PCM pulses 221-226 from the respective adders 351-355 and suppress the finite output where unnecessary.

Referring to FIGS. 20 and 21, the signal combiner 260 comprises a first through a sixth channel. The jth channel shown in FlG. 20 comprises a first and a second NAND gate 391 and 392 supplied with the jth digits of the first delayed and the mean PCM pulses 231-236 (P10. 21) and 251-256 (reproduced in FIG. 21), respectively, an AND/NAND gate 393 supplied with the outputs of the NAND gates 391 and 392, a first flip-flop 396 supplied with the AND and the NAND outputs of the AND/NAND gate 393 and stepped by the fifth clock pulses 207 (FIG. 21), and a second flip-flop 397 supplied with the AND and the NAND outputs of the first flipflop 396 and stepped by the sixth clock pulses 208 (FIG. 21 The NAND output a (FIG. 21) of the AND/NAND gate 393 is given by the following logic relation:

where ri and d represent the jth digits of the PCM pulses 231-236 and 251-256, respectively. This relation shows that the mean PCM pulses 251-256 are superposed on the first delayed PCM pulses 231-236.

It should be mentioned here thatthe utilization circuit 270 may comprise a decoder for decoding the reproduced PCM pulses 2611 266 and further a low-pass filter of the same cutoff frequency as the low-pass filter on the transmitter side, for obtaining the functional frequency component of the decoded analogue signal.

Referring finally to FIGS. 22 and 23A-D another receiver for use together with the transmitter of this invention comprises an input terminal 200, a clock generator 201, a seriesparallel converter 210, and a time rearrange circuit 220, all similar to the corresponding components explained in conjunction with the receiver of F IO. 11.

The receiver of FIG. 22 further comprises a decoder 228' for decoding the time rearranged PCM pulses 221-226 to produce time rearranged samples 229' (FIG. 23A), a first delay line 230' forgiving a time delay of the sampling period t to the time-rearranged samples 229' to produce first delayed samples 239', a second delay line 240 also having a delay time oft for similarly producing second delayed samples 249', an adder 250' for adding the time rearranged samples 229' and the corresponding second delayed samples 249 to produce sum samples, a divider 250" for dividing the sum samples by two to produce mean samples 258' (FIG. 23B), and a spurious mean sample suppressor 250" controlled by the variation-representing seventh-digit pulses 217 for suppressing spurious mean samples contained in the mean samples to provide interpolation samples 259' (FIG. 23C) More particularly, the suppressor 250' suppresses in this case the spurious samples 8,, 8,, S 8., 8,, and S, and produces the interpolation samples S, and 8,.

The receiver of FIG 22 still further comprises a signal combiner 260' for superposing the interpolation samples 259' on the first delayed samples 239' to produce approximately reproduced samples 269' (FIG. 23D). The receiver preferably comprises a resampling circuit 401 for resampling the approximately reproduced samples 269' with a view to eliminating the irregularities introduced into the shapes of the signal to result in noises by various time delays used during the signal processing. The receiver may comprise a low-pass filter 402 for deriving the fundamental frequency band 403 (FIG. 23D).

It will now be apparent to those skilled in the art that it is possible to adapt the parabolic interpolation explained with reference to FIGS. 6 and 7 instead of the linear interpolation illustrated throughout the description of the transmitter and the receiver. Also, it is possible to apply the invention to transmission of multiplex PCM signals. Either analogue or digital processing of the signals may be employed in conformity with the nature of the original signal, the accuracy and stability required, the simplicity of the circuit, and other factors.

It is to be noted that the value of the threshold signal It serves as a parameter for determining without any ambiguity the error between the original data and the approximately reproduced data for given number of bits in each original PCM codeword, number of bits in each PCM codeword to be transmitted, and manner of interpolation. It is therefore possible to determine the optimum value of the threshold signal It by calculating the solution of the equation for such error under a certain criterion, such as the least squares. Inasmuch as the application of this invention to the transmission of television picture signals is based on psychophysics, the criterion should then depend on the psychologic measure and consequently the optimum threshold value II should be determined through experiments carried out from a subjective vrew.

What is claimed is:

1. A transmitter for sending an analog signal varying in magnitude as a pulse code modulation signal compressed in band width, comprising:

a source of an analog signal varying in magnitude;

a generator of timing signals having different repetitive times;

means activated by successive clock signals occuring at a repetitive sampling time t for encoding one portion of said analog signal into a first preselected number of first sets, each consisting of a first preassigned number of parallel pulse code modulation pulses containing most significant digits and least significant digits and corresponding to one sample of said analog signal;

means for compressing said first sets in bandwidth;

means for transmitting said first sets as compressed in bandwidth; and

means for activating said compressing means to control the bandwidth compression of said first sets and simultaneously therewith to sense the speed of variation in magnitude between successive first sets and thereby between corresponding successive analog signal samples, includmg:

means activated by a second portion of said analog signal and a further timing signal occurring at a repetitive time 2! for producing "l" and logic signals to indicate rapid and slow rates of speed variation, respectively, in the magnitude of said analog signal at successive even number first sets and thereby at corresponding successive even number analog signal samples; and

circuit means for applying to said compressing means said 1" and 0" logic signals and a predetermined number of additional timing signals of which one occurs at said sampling time I and others occur at a time 2: to compress said first pulse sets into a preselected number of second sets of parallel pulse code modulation pulses; said second preselected number being less than said first preselected number;

whereby each set of said second sets is provided with said more significant digits of one odd number of said first sets; certain sets of said second sets are provided with said more significant digits thereof combined with said less significant digits of corresponding odd number first sets as slow speed variations occur in the magnitude of said analog signal at even number first sets next adjacent to said respective last-mentioned odd number first sets; other sets of said second sets are provided with said more significant digits of corresponding other odd number first sets combined with said more significant digits of other even number first sets as rapid speed variations occur in the magnitude of said analog signal at other even number first sets next adjacent to said last-mentioned other odd number first sets; and final digits comprising 0" and "l" logic signals are provided in said certain and other second sets to indicate said slow and rapid variations, respective ly, in said analog signal magnitude at said respective even number first sets.

2. The transmitter according to claim 1 in which said control means includes means responsive to said analog signal second portion for deriving therefrom a signal varying in magnitude and representing a varying difierence in magnitude between a true value and a calculated value of said last-mentioned portion, thereby to sense said rapid and slow variations in said analog signal magnitude.

3. The transmitter according to claim 2 in which said control means includes:

a source of threshold voltage having a fixed magnitude; and

means for comparing the magnitude of said derived difference voltage and said threshold voltage magnitude at successive times 21 corresponding to said first-mentioned and other even number first sets to produce said 0 and l logic signals.

4. A transmitter for an analog signal varying in magnitude as a pulse code modulation signal compressed in bandwidth, comprising:

a source of an analog signal varying in magnitude;

a generator of timing signals having different repetitive times;

means activated by predetermined number of successive timing signals occurring at repetitive sampling times t for encoding one portion of said analog signal into a preselected number of first sets, each consisting of a first preassigned number of parallel pulse code modulation pulses containing more significant digits and less significant digits and corresponding to one sample of said analog signal;

means for compressing said first sets in bandwidth;

means for transmitting said first sets as compressed in bandwidth; and

means for activating said compressing means to control the bandwidth compression of said first sets and simultaneously therewith to sense the speed variation in magnitude between successive first sets and thereby between corresponding successive analog signal samples, including: means responsive to a second portion of said analog signal for deriving therefrom a signal varying in magnitude and representing a varying difference in magnitude between a true value and a calculated value of said last-mentioned portion thereby to sense the speed variation in the magnitude of said analog signal at successive first sets and thereby at corresponding successive analog signal samples;

a source of threshold voltage of fixed magnitude;

means for comparing the magnitudes of said derived voltage and said threshold voltage at successive timing signals 2! corresponding to successive even numbers of said first sets to produce two different logic signals to represent the speed variation in the magnitude at said last-mentioned even number first sets, a first of said two logic signals representing a slow speed variation magnitude and a second of said two logic signals representing a rapid speed variation magnitude, each of said one and other logic signals having a time duration 21;,

means for delaying said first and second logic signals to register in time with two successive first sets, each including one of said even number first sets; and

circuit means for applying to said compressing means and first and second speed variation logic signals and a predetermined number of additional timing pulses of which one occurs at said sampling time t and others occur at a time 2! to compress said first sets into a preselected number of second sets, each consisting of a second preassigned number of parallel pulse code modulation pulses and one of said first and second logic'signals; said second preselected number being less than said first preselected number;

whereby each of said second sets includes said more significant digits of one odd number of said first sets; each of said last-mentioned digits doubled in width and delayed by one sampling time t relative to said respective first and second speed-variation logic signals; certain sets of said second sets are provided with said more significant digits thereof combined said less significant digits of corresponding odd number first sets as slow speed variations occur in the magnitude of said analog signal at even number first sets adjacent to said last-mentioned odd number first sets as indicated by said first logic signals; each of said last-mentioned less significant digits doubled in width and delayed by one sampling time t; and other sets of said second sets are provided with said more significant digits of corresponding other odd number first sets combined with said more significant digits of other even number first sets as rapid speed variations occur in the magnitude of said analog signal at other even number first sets next adjacent to said last-mentioned odd number first sets; and final digits comprising 0" and 1" logic signals corresponding to said first and second speed variation signals, respectively, are provided in said certain and other second sets to indicate said respective slow and rapid variations in said analog signal magnitude at said respective even number first sets.

5. A transmitter for sending an analog signal varying in magnitude as a pulse code modulation signal, comprising:

a source of analog signal varying in magnitude;

means for encoding a first portion of said analog signal into a preselected number of first sets of parallel pulse code modulation pulses, each set representing a corresponding sample of said analog signal and including a first preassigned number of said parallel pulses including more significant digits and less significant digits;

means activated by a second portion of said analog signal for compressing said first pulse sets into a second preselected number of second sets of parallel pulse code modulation pulses; said second preselected number being smaller than said first preselected number; each set of said second pulse sets including said more significant digits of one odd number set of said first sets; certain sets of said second pulse sets having said more significant digits thereof combined with said less significant digits of corresponding odd number first sets as slow speed variations occur in the magnitude of said analog signal at even Li number first sets next adjacent to said respective lastmentioned odd number first sets; other sets of said second sets having said more significant digits of corresponding other odd number first sets combined with said more sig- 4t, a second timing pulse recurring at a time 2!, and third and fourth timing pulses, each recurring at a time t;

means activated by said first timing pulses for converting said source output preselected second set series pulse code modulation pulses into a preselected number of third sets of pulse code modulation pulses of combined parallel more significant digits, less significant digits, and and l logic signals corresponding to said combined more significant digits, less significant digits, and said 0 and 1" signals, respectively, of said preselected second pulse sets;

means responsive to said second, third and fourth timing greater than said preselected first number; certain sets of said fourth sets provided with combined parallel more significant digits and less significant digits corresponding to said odd number first pulse sets consisting of more signifinificant digits of other even number first sets as rapid 5 cant digits and less significant digits during first halves of speed variation occurs in the magnitude of said analog said respective second timing pulses 2t and followed by signal at other even number first sets next adjacent to said sets of parallel 0" logic signals during the second halves last-mentioned other odd number first sets; and a final of said last-mentioned pulses when said final pulses in said digit included in each of said certain second sets and comrespective third pulse sets are "0" logic signals; and other prising a 0" logic signal to indicate said slow variation sets of said fourth pulse sets provided with more signifianalog signal magnitude at each of said first-mentioned cant digits corresponding to said more significant digits of even number first sets and a 1" logic signal in each of other odd number first pulse sets during first halves of said other second sets to indicate said rapid variation said respective second timing pulses 2t and followed by analog signal magnitude at each of said other even 1 5 more significant digits corresponding to said more signifinl-lmbcr firs! acts; and cant digits of said other next adjacent even number first means for transmitting said second sets of parallel pulse pulse t; d i seoond halves f said last-mentioned code modulation pulses pulses when said final pulses in said respective other third 6. A receiver for a pulse code modulation signal comprised sets are l i 1" i l said parallel more significant in bandwidth to representananalog signal, comprising: di i i id respective other f r h sets f ll d b a source of output series pulse code modulation pulses coma ll l sequential sets of "1," 0" and 1" logic signals;

pressed in bandwidth to represent an analog signal varya d ing in magnitude and divided into a preselected number means for utilizing said fourth pulse sets to reproduce said of first sets of parallel pulse code modulation pulses, each analog signal varying in magnitude. set representing a corresponding sample of said analog 5 7. The receiver according to claim 6 in which said utilizing signal and including a first preassigned number of said means i el d parallel pulses consisting of more significant digits and means for delaying a first portion of said fourth pulse sets by less significant digits; said pulses of said first sets comfi t 1 digit ti e interval; and pressed into a preselected number of second sets of paralmean for delaying a first portion of said l-digit delayed first lel pulse code modulation pulses; said preselected first set portion of said fourth pulse sets by a second l-digit time number being greater than said preselected second set i terval. number; each set of said second sets including said more 8. The receiver according to claim 7 in which: significant digits of one number set of said first sets; cersaid timing means generates a fifth timing pulse recurring at tain sets of said second sets having said more significant a time 2!; and digits thereof combined with said less significant digits of said utilizing means comprises means responsive to said 0" corresponding odd number first sets as slow speed variaand 1" signals included in said third pulse sets and to tions occur in the magnitude of said analog signal at next said fifth timing pulse for digitally adding a second poradjacent even number first sets; other sets of said second tion of said fourth pulse sets and said first portion of said sets including said more significant digits of correspondfourth pulse sets delayed by said first and second l-digit ing other odd number first sets combined with said more time intervals and thereafter dividing such digital sum by significant digits of next adjacent other even number first two to produce a fifth preselected number of means sets sets as rapid speed variations occur in the magnitude of of parallel pulse modulation pulses, said first and fifth said analog signal at said last-mentioned sets; and final preselected numbers being identical; said fifth pulse sets digits included in said respective second sets and compriscomprising sets of 0" logic signals corresponding to said ing 0" and 1" logic signals to indicate said slow and certain and other sets included in said fourth pulse sets rapid variation analog signal magnitudes, respectively, at and further sets of combined more significant and less sigeach of said first-mentioned even number first sets; each nificant digits corresponding to sets of parallel 0" signals of said digits in each of said second sets and said respecincluded in said fourth pulse sets. tive 0 and "1 signals having a time duration of 2!, t 9. The receiver according to claim 8 in which: being the time of each of said analog signal samples; leadsaid timing means generates sixth and seventh timing pulses, ing edges of said second set digits including said comeach recurring atatime t; bined more significant and less significant digits corsaid utilizing means comprises means responsive to said responding with leading edges of associated 0 signals sixth and seventh timing pulses for superposing said fifih and leading edges of said second set digits including said pulse sets on a second portion of said l-digit delayed first combined more significant digits corresponding with portion of said fourth pulse sets to produce a sixth leading edges of associated 1" signals; said second set preselected number of sets parallel pulse code modulaparallel pulse digits present at said source output as cortion pulses; said first and sixth preselected numbers being responding series pulse code modulation pulses; identical; certain sets of said sixth pulse sets comprising means for generating a first timing pulse recurring at a time combined more significant digits and less significant digits corresponding to said combined more significant digits and less significant digits included in said certain fourth pulse sets, first additional sets of combined more significant digits and less significant digits corresponding to said combined more significant digits and less significant digits included in said fifth pulse sets, and second additional sets of combined more significant digits and sequential"l," 0" and 0" logic signals corresponding to said respective other fourth pulse sets including combined more significant digits and sequential l," 0" and 0" logic signals.

10. The receiver according to claim 9 in which said utilizing pulses for rearranging said third pulse sets to provide a preselected number of fourth sets of parallel pulse code modulation pulses; said preselected fourth number being means includes a utilizing circuit for translating said sixth pulse sets into a reproduction of said analog signal varying in magnitude.

i ll? ll. The receiver according of claim ti in which said utilizing means includes:

means for decoding said rearranged fourth pulse sets to provide time rearranged fifth pulse sets corresponding to said certain and other sets of said fourth sets and to omit pulses corresponding to said sets in said fourth sets; and

means for translating said fifth pulse sets to reproduce said analog signal varying in magnitude.

12. The receiver according to claim ill in which said translating means includes:

means for delaying a first portion of said fifth pulse set by a first l-digit time interval;

means for delaying a second portion of said l-digit delayed first portion of said fifth pulse sets;

means for adding said first and second l-digit delayed first portion of said fifth pulse sets and a second portion of said fifth pulse sets to produce sum samples thereof;

means for dividing said sum samples to produce mean pulse samples;

means responsive to said 0" and 1" signals included in said third pulse sets for converting said mean pulse samples into interpolation pulse samples to represent said omitted pulses in said fifth pulse sets;

means for combining a second portion of said l-digit delayed first portion of said fifth pulse sets and said interpolation pulse samples to produce a sixth set of pulses whose envelope represents the envelope of said analog signal; and

means for utilizing said sixth pulse set to reproduce said analog signal.

13. A system for transmitting and receiving an analog signal varying in magnitude as a pulse code modulation signal compressed in bandwidth, comprising;

a transmitter consisting of:

a source of an analog signal varying in magnitude;

means for encoding a first portion of said analog signal into a preselected number of first sets of parallel pulse code modulation pulses, each set representing a corresponding sample of said analog signal and including a first preassigned number of said parallel pulses including more significant digits and less significant digits;

means activated by a second portion of said analog signal for compressing said first pulse sets into a preselected number of second sets of parallel pulse code modulation pulses; said preselected second number being smaller than said preselected first number; each set of said second pulse sets including said more significant digits of one odd number set of said first sets; certain sets of said second pulse sets having said more significant digits thereof combined with said less significant digits of corresponding odd number first sets as slow speed variations occur in the magnitude of said analog signal at even number sets next adjacent to said respective last-mentioned odd number first sets; other sets of said second sets having said more significant digits of corresponding other odd number first sets combined with said more significant digits of other even number first sets as rapid speed variations occur in the magnitude of said analog signal at said last-mentioned sets; and final digits included in said certain second sets and comprising 0" logic signals to indicate said slow variation analog signal magnitude at said respective firstmentioned even number first sets and 1" logic signals to indicate said rapid variation analog signal magnitude at said respective other even number first sets; each of said digits in each of said second sets and each of said 0" and 1" signals having a time duration 2!, I being the time duration of each of said analog signal samples; leading edges of said digits of said certain second sets corresponding with leading edge of associated ll signals and leading edges of said digits of said other second sets corresponding with leading edges of as sociated ll signals; and

means for transmitting said second pulse sets as series pulse code modulation pulses; and a receiver consisting of:

means for receiving said transmitted series pulse code modulation pulses and converting said last-mentioned pulses into a preselected number of third sets of pulse code modulation pulses; said preselected third number and said preselected second number being identical; each of said third sets including combined parallel more sigrificant digits, less significant digits, and 0" and 1" logic signals corresponding to said combined more sigrificant digits, less significant digits and 0" and 11" logic signals included in each of said second sets; and means activated by said third pulse sets for providing a reproduction of said analog signal varying in amplitude. 11d. A system for transmitting and receiving an analog signal varying in magnitude as pulse code modulation signal compressed in bandwidth, comprising:

a transmitter consisting of:

a source of analog signal varying; in magnitude; means for encoding a first portion of said analog signal into a preselected number of first sets of parallel pulse code modulation pulses, each set representing a corresponding sample of said analog signal and including a first preassigned number of said parallel pulses including more significant digits and less significant digits; means activated by a second portion of said analog signal for compressing said first pulse sets into a preselected number of second sets of parallel pulse code modulation pulses; said preselected! second number being smaller than said preselected first number; each set of said second pulse sets including said more significant digits of one odd number set of said first sets; certain sets of said second sets having said more significant digits thereof combined with said less significant digits of corresponding odd number first sets as slow speed variations occur in the magnitude of said analog signal at even number sets next adjacent to said respective last-mentioned odd number first sets; other sets of said second sets having said more significant digits of corresponding other odd number first sets combined with more significant digits of other even number first sets as rapid speed variations occur in the magnitude of said analog signal at said last-mentioned sets; and final digits included in said certain second sets and comprising 0" logic signals to indicate said slow variation analog signal magnitude at said first-mentioned even number first sets and 1" logic signals to indicate said rapid variation analog signal magnitude at said respective other even number first sets; each of said digits in each of said second sets and each of said 0" and l signals having a time duration 2!, I, being the time duration of each of said analog samples; leading edges of said digits of said certain second samples corresponding with leading edges of associated (0" signals and leading edges of said digits of said other second sets corresponding with leading edges of associated l" signals; means for transmitting said second pulse sets as series pulse code modulation pulses; and a receiver consisting of:

means for receiving said transmitted series pulse code modulation pulses and converting said last-mentioned pulses into a preselected number of third sets pulse code modulation pulses corresponding to said preselected number of second sets of pulse code modulation pulses; means for rearranging said third pulse sets to provide a preselected number of fourth pulse sets of parallel pulse code modulation pulses; said preselected fourth number being greater than said preselected third number; certain sets of said fourth sets provided with more significant and less significant digits corresponding to said odd number first sets consisting of more significant digits and less significant digits and followed by sets of parallel logic signals when said final pulses in said third pulse sets are 0" logic signals; and other sets of said fourth sets provided with more significant digits corresponding to more significant of other odd number first sets followed by more significant digits corresponding to said more significant digits of said other next adjacent even number first sets where said final pulses in said respective other third sets are 1" logic signals; said more significant digits in said respective other fourth sets followed by parallel sequential all sets of l," 0" and 0" logic signals;

means for decoding said rearranged fourth pulse sets to provide time rearranged fifth pulse sets corresponding to said certain and other sets of said fourth sets and to omit pulses corresponding to said 0" sets in said fourth pulse sets;

means for converting said fifth pulse sets into a sixth set of pulses whose envelope represents the envelope of said transmitter analog signal; and

means for utilizing said sixth pulse set to reproduce said transmitter analog signal.

t i l l

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US2722660 *Apr 29, 1952Nov 1, 1955Jones Jr John PPulse code modulation system
US3097338 *Sep 26, 1961Jul 9, 1963Eugene Pinet AndrePulse-code modulation transmission systems
US3330909 *Jan 2, 1964Jul 11, 1967Bell Telephone Labor IncPulse communication system
US3393364 *Oct 23, 1965Jul 16, 1968SignatronStatistical delta modulation system
US3398239 *May 21, 1964Aug 20, 1968IttMultilevel coded communication system employing frequency-expanding code conversion
US3502806 *Aug 1, 1966Mar 24, 1970Xerox CorpModified run length data reduction system
US3505470 *Mar 28, 1967Apr 7, 1970IbmProcess and device for coding and decoding digital signals via phase modulation
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US4071842 *Jan 3, 1977Jan 31, 1978Bell Telephone Laboratories, IncorporatedApparatus for analog to digital conversion
US4125861 *Aug 18, 1977Nov 14, 1978Bell Telephone Laboratories, IncorporatedVideo signal encoding
US4412349 *Oct 19, 1981Oct 25, 1983Delta Electronics, Inc.Remote display of percentage modulation data obtained at a transmitter site
US4553249 *Oct 4, 1983Nov 12, 1985Siemens AktiengesellschaftMethod and circuit arrangement for converting a binary signal, alternating between two levels, to a pulse code signal which comprises data pulses and renewal pulses
US4558455 *Sep 23, 1983Dec 10, 1985International Business Machines CorporationData transmission system
US4602343 *Mar 19, 1982Jul 22, 1986Mcc PowersSupervisory and control system having dynamic significant change of values
US4680797 *Jun 26, 1984Jul 14, 1987The United States Of America As Represented By The Secretary Of The Air ForceSecure digital speech communication
US5659660 *Mar 22, 1993Aug 19, 1997Institut Fuer Rundfunktechnik GmbhMethod of transmitting and/or storing digitized, data-reduced audio signals
US5737367 *Oct 7, 1994Apr 7, 1998U.S. Philips CorporationTransmission system with simplified source coding
Classifications
U.S. Classification375/242, 375/241, 375/E07.26, 704/212
International ClassificationH04B1/66, H04N7/12, H04B14/04, H04N7/26
Cooperative ClassificationH04B14/046, H04N19/00
European ClassificationH04N7/26, H04B14/04D
Legal Events
DateCodeEventDescription
Jul 30, 1985ASAssignment
Owner name: NIPPON TELEGRAPH & TELEPHONE CORPORATION
Free format text: CHANGE OF NAME;ASSIGNOR:NIPPON TELEGRAPH AND TELEPHONE PUBLIC CORPORATION;REEL/FRAME:004454/0001
Effective date: 19850718